Fire Risk Management using Data Cubes, Machine Learning and OBDA systems.

We present a fire risk management system which takes input data from various sources (e.g., meteorological data, satellite indicators for vegetation, historical burned areas), produces a harmonized spatio-temporal data cube to compute fire risk and enables semantic querying to assist fire risk management. The distinguishing implementation features of the system is the use of data cubes, machine learning algorithms and, most importantly, geospatial ontology-based data access technologies. The system has been implemented in the European project DeepCube for the geographic area of Greece and can be used operationally to assist authorities to determine fire risk during the summer fire season.